Computational Science Technical Note CSTN-090


Elucidating Soldier and Worker Caste Divisions in an Animat Artificial Life Model

C. J. Scogings and K. A. Hawick

Archived May 2009, Revised March 2010


Complex systems such as termite colonies have a macroscopically rich set of emergent behaviours that are thought to emerge solely from microscopic and simple individual agent behaviours. The effect of different castes of termite on the success of a system as a whole is a subtle effect that is difficult to describe analytically, but which can be explored using simulation techniques. We explore the effect of differentiating worker and soldier castes in a predator-prey based animat-agent model. We describe our simulation model and its implementation. The core model comprising a well-established set of microcscopic rules and behavioural priorities normally exhibits statistically predictable boom-bust phenomena with population booms of the predators lagging those of prey. Introducing soldier and worker castes amongst predators can exacerbate and exaggerate the normal model properties. We present results showing the relative advantages to a termite community of having varying proportions of soldiers to worker and we also discuss resource utilisation efficiency and sociological implications of this mode.

Keywords: artificial life; agent technologies; castes; termite colony; animat agents; bio-inspired systems.

Full Document Text: PDF version.

Citation Information: BiBTeX database for CSTN Notes.

BiBTeX reference:

  author = {C.J. Scogings and K.A. Hawick},
  title = {Elucidating Soldier and Worker Caste Divisions in an Animat Artificial
	Life Model},
  booktitle = {International Conference on Genetic and Evolutionary Methods (GEM'10)},
  year = {2010},
  pages = {187-193},
  address = {Las Vegas, USA},
  month = {12-15 July},
  publisher = {WorldComp},
  note = {GEM4365},
  timestamp = {2010.05.02}

[ CSTN Index | CSTN BiBTeX ]